Introduction

Patients with immune-mediated rheumatic diseases (IMRD) have higher risk of infections, particularly associated with immunossuppresion and disease activity. However, patients with spondyloarthritis (SpA) have no increased risk for severe infection, even after exposing to the TNF (TNFi) and IL17 inhibitors (IL17i) [1], differently of individuals with systemic erythematous lupus (SLE) and rheumatoid arthritis (RA) [1,2,3,4]. Some aspects could be highlighted to try to explain these findings when the IMRDs are compared, including lower use of glucocorticosteroids and other immunosuppressive drugs [Statistical analysis

To characterize the patient profile, the frequencies percentage and mean and standard deviation (SD) of variables were calculated. Comparisons of means between two groups were performed using the Student’s t test for independent samples. To verify normality data, it was applied the Kolmogorov-Smirnov test. In case normality violation, it was used the Mann-Whitney non-parametric test. The chi-square association test was used to assess the association among categorical variables with standardized adjusted residual calculation, or Fischer’s exact test for small samples. The final logistic regression model used moderate/ severe forms as dependent variable and appropriate adjustments were performed considering all independent variables that had statistical significance up 10% in the univariate analysis. P value was set as significant if below 5%. Statistical analyzes were performed using the SPSS 20.0 statistical software.

Results

A total of 153 axial SpA patients were included, of whom 85 (55.5%) with COVID-19 and 68 (44.4%) without COVID-19. Most of them were men (N = 92; 60.1%) with mean age of 44.0 ± 11.1 years and long-term disease (11.7 ± 9.9 years), and 112 (73.2%) patients had positivity status for HLA-B27. All patients included were COVID-19 confirmed cases, according to Brazilian Ministry of Healthy recommendations, most of them classified according to lab criterion (80.1%), mostly through RT-PCR (58.5%). Regarding biological DMARDs, 110 (71.8%) were on TNF inhibitors and 14 (9.15%) on IL-17 antagonists. No significant difference was found regarding concomitant medications, including conventional or biologic DMARDs between the groups There were no significant statistically differences concerning social distancing, smoking, BMI, waist circumference and comorbidities. Also, there were no significant difference concerning ASDAS-CRP or ASDAS-ESR, but SpA patients HLA-B27 negative had higher score related to pain and other PROs (patient-reported outcomes) measured though the VAS and BASDAI (Table 1). However, when stratifying the relationship between the disease activity, measured by these variables (VAS and BASDAI), and the COVID-19 severity, no significant differences were observed (Table 2). Interestingly, no new episodes of arthritis, enthesitis or extra-musculoskeletal manifestations were reported after the COVID-19.

Table 1 Clinical and demographic data of the sample, comparing HLA-B27 positive with HLA-B27 negative patients
Full size table
Table 2 Disease activity, measured by VAS and BASDAI, and the COVID-19 severity in SpA patients
Full size table

Comparing SpA patients with and without COVID-19, the HLA-B27 positivity was not different between groups (n = 64, 75.3% vs. n = 48, 48%, respectively; p = 0.514). Similarly, when comparing mild, moderate, and severe COVID-19, according to the HLA-B27 status, no significant difference was found (Table 3). The mean time from the first symptoms to hospitalization was 7.1 ± 3.4 days, and although the number of hospitalization days has been numerically higher in the B27 positive group than in those negative status, no statistically significant difference was observed (13.5 ± 14.8 vs. 5.7 ± 4.11, respectively; p = 0.594). Only one HLA-B27 negative patient died. The hospitalization frequency, considered as moderate COVID, was higher in the B27 positive patient group (p = 0.031) (Table 3). Disease activity was similar before and after the infection the control group had significantly higher disease activity score, according to ASDAS-CRP (3.8 vs. 3.2, p = 0.08).

Table 3 Outcomes related to COVID-19 (only case group), according to the HLA-B27 status in spondyloarthritis patients
Full size table

Discussion

Our data demonstrated no significant difference of COVID-19 frequency rate or severity pattern in patients with axial SpA regarding the HLA-B27 positivity status, suggesting a lack of protective effect against the SARS-CoV-2 infection. In addition, the disease activity was similar before and after the infection. Our data emphasize that the disease behavior in axial SpA patients is similar to the general population, regardless biologic therapy and HLA-B27 status.

Similarly, Rosenbaum et al. using an online self-reported survey in 3,435 SpA patients from 65 countries (85% with AS; 76.1% were HLA-B27 positive from 2,836 aware of HLA-B27 status) from the Spondylitis Association of America (SAA) between 10 and 2020 and 26 April 2021, showed also similar rate of infection between HLA-B27 positive (1.4%) and negative (1.5%), as well as only in those tested (0,6% in both groups) and without significant difference related to self-reported COVID-19 severity. In addition, they did not find any significant difference the incident rate ratio for COVID-19 compared to the US population adjusted for age and sex, regardless HLA-B27, diagnosis of SpA or its treatment. On the other hand, it is important to note that some potential bias, such as responders’ profile, low number of infected individuals (n = 41), self-reported diagnosis of SpA and HLA-B27 status, as well as COVID-19, including confirmatory lab testing in only 36.9% of sample and severity based on patient’s opinion), and some limitations, such as lack of information regarding the disease activity data. Thus, our prospective cohort study provides more consistent data regarding the relationship between HLA-B27 and COVID-19 susceptibility and severity [33,34,35].

During human evolution, the existence of conserved HLA haplotypes is resulting from a selective pressure by some pathogens that have edited the immune response genes and provided higher survival probability over time. Considering that the virus-specific T cell response is critical for determining the infection severity, as well as viral clearance, interferon release, duration of immunity and efficacy of vaccines, our initial hypothesis was driven to the protective role of HLA-B27 against SARS-CoV-2 infection in axial SpA patients, a similar finding found in other viral infections. However, our hypothesis was not confirmed in this prospective face-to-face analysis [36].

Regarding specifically the SARS-CoV-2, the infectivity, severity, and mortality rates related to COVID-19 have not affected all countries equally, suggesting that several aspects could be involved in different populations. Beyond traditional risk factors, such as age, comorbidities, lack of masking/ social distancing and viral load, some authors highlight the relevance of genetic biomarkers as potential drivers for ethnic susceptibility and unfavorable prognosis [37, 38]. Although the HLA systems have a strong linkage disequilibrium that can result in conserved multi-locus haplotypes depending on how the alleles are arranged in adjacent loci of a specific region of the chromosome, it is the best way to characterize the host genetic variance concerning the innate immunity, especially related to MHC class I, which plays a major role in susceptibility to viral infections [39]. Thus, these aspects may contribute to the difference in mortality among various countries [40]. However, a recent review found heterogeneous findings in the association of inter-individual genetic variants with COVID-19 susceptibility and severity COVID-19, but without any relationship with HLA-B27 [41].

Furthermore, NK and dendritic cells, neutrophils and other antigen presenting cells with different signaling ways, including PD-1 and cytokine storm, and exhausted lymphocytes are involved in the COVID-19 pathophysiology [36]. Recently, it was demonstrated CD8+ T cells derived from SARS-CoV-2 survivors exhibited polyfunctional effector responses to two novel NC-derived peptides identified as HLA-binders, although not specified [36, 42, 43].

Considering previous outbreaks, Lin et al. [44] reported higher risk and severity of SARS-CoV infection in health care workers carrying HLA-B*4601 (risk and severity) and HLA-B*5401 (only risk) in Taiwan [45]. On the other hand, no significant association was observed in Chinese individuals regarding genotypic patterns of HLA-A, -B and -DRB1 loci in SARS patients when compared to a co-resident population. In Saudi individuals, Hajeer at al demonstrated significant association with MHC class II (HLA-DRB1*11:01 and DQB1*02:02) and susceptibility to severe MERS-CoV (Middle East respiratory syndrome) [45, 46].

Regarding the COVID-19, Leite et al. evaluated data sets of HLA-B alleles, KIR genes and functional single nucleotide polymorphisms (SNPs) in cytokines related to COVID-19 cytokine storm and found significant correlation between eight HLA-B alleles and polymorphisms in three cytokine genes (IL6, IL10, and IL12B) and daily death rates across countries [47]. Sakuraba et al. demonstrated significant association between HLA-C*05 and higher COVID-19 mortality in 74 countries, after multiple adjustments. A 1% increase in the allele frequency of HLA-C*05 was associated with an increase of 44 deaths/million. Noteworthy, the combination of HLA-C*05 to its receptor KIR2DS4fl, expressed on NK cells, causes NK cell-induced hyperactive immune response and higher mortality rate. Countries with similar ethnic and/or geographic background responded in a similar pattern to each pandemic [48].

Although the HLA genotypes might affect the susceptibility to SARS-CoV-2 infection or severity of COVID-19, these aspects are not still clear up to present day. Toyoshima et al. identified a total of 1234 mutations by comparing with the reference SARS-CoV-2 sequence after analyzing 12,343 SARS-CoV-2 genome sequences of individuals from 28 countries in six geographic areas. They found that ORF1ab 4715 L and S protein 614G variants and the frequency of several HLA alleles, including HLA-A*11:01, were associated with number of infected cases and not as an independent factor for fatality rate in each country [49].

More recently, Pretti et al., evaluating 140 HLA alleles and the landscape of 3,723 potential HLA-I A and B restricted SARS-CoV-2-derived antigens in 37 countries, added a new information regarding three potential antigens coverage of S and N derived peptides and the number of deaths [50]. Similarly, other authors have also reported same findings related to SARS-CoV-2 immune responses, susceptibility and different genetic backgrounds. Interestingly, Guo et al. found diverse capacities of S protein specific epitope presentation by different HLA alleles with very limited number of predicted epitopes for HLA-B*2705, HLA-B*4402 and HLA-B*4403 and as high as 132 epitopes for HLA-A*6601, suggesting individuals with HLA-B*44 are more prone to SARS-CoV-2 infection. These findings could help us to better understanding the role of HLA-B27 in the COVID-19 [25, 51].

Thus, some points could be highlighted to explain the lack of association between COVID-19 and HLA-B27. Firstly, only Guo study showed some relationship with B27 in general population database across the countries. Secondly, the cytokines pattern and inflammatory signaling ways involved with severe COVID do not seem to be related to Th17 activation, an immune-inflammatory via strongly associated with SpA. Thirdly, the low plasticity of epitopes found by Guo et al. in HLA-B27 could not be crucial to influence the antigen presentation to the spike proteins from SARS-CoV-2 and other coronavirus family. Fourthly, severe COVID-19 have been reported in all IMRDs and the unfavorable outcome has been more related to age and comorbidities than underlying rheumatic disease. On the other hand, the SARS-CoV-2 could work as a trigger to a myriad of musculoskeletal syndromes. Although some authors have reported anecdotal cases of reactive arthritis after COVID-19, the most recent systematic review showed that the manifestations with consistent association with SARS-CoV-2 infection were autoimmune cytopenia, cutaneous vasculitis, encephalitis, and Guillain-Barre syndrome [52]. Considering the inclusion period ended before the massive vaccination against the COVID-19, it is important to mention that our data are related to the SARS-CoV-2 itself and HLA-B27 positivity without any interference of adenovirus vectors, inactivated full virus or mRNA vaccines.

Another important thing to be considered is related to endoplasmatic reticulum aminopeptidases (ERAPs) [53]. As they are pivotal enzymes for trimming peptides and to generate optimal-length antigens to fit into the MHC class I groove, the SARS-CoV-2 likely must not cause any important impairment in this cell step neither related to repertoire of antigens presented by HLA-B27. The inadequate activation of NK and CD8+ T cells could have other pathophysiological mechanisms. Nonetheless, these points are merely speculative because there is no information on the ERAP1, ERAP2, and LNPEP, as well as co-segregation with HLA-B27 subtypes, which are involved in antigen presentation and peptide trimming and the COVID-19 occurrence and severity. Interestingly, it is important to note these multifaceted aminopeptidases regulates the renin–angiotensin system and higher susceptibility to hypertension, other aspects associated with severe COVID-19 [54]. Considering that an increased risk of some comorbidities has been reported in patients with spondyloarthritis, including hypertension and other cardiovascular diseases, more attention could be given to these aspects to understand better this complex interaction among genetic, environmental and several cellular processes [55].

Additionally, the concomitant conventional and biological DMARDs are also related to higher risk of infections, particularly related to immune response impairment against several pathogens, we did not find any significant association with severe COVID-19 in our SpA database, regardless HLA-B27 status. Considering all forms of COVID-19, we observed 19 cases in B27 negative SpA patients and 52 in those B27 positive, suggesting lack of protective role regarding the SARS-CoV-2 infection. However, when evaluating only hospitalized patients (moderate COVID-19), there was a greater number of patients in the B27 negative group. It is important emphasizing that the effect size was very small (OR = 0.13; 95%CI 0.02–0.81) in only 6 patients (4 HLA-B27 negative and 2 HLA-B27 positive), which does not allow us to draw conclusions about that.

To our best knowledge, our database is the largest and robust database regarding SpA patients during COVID-19 pandemic and comparing to SpA patients no infected by SARS-CoV-2. However, our ongoing observational prospective registry has some limitations, such as short period of follow-up (until infection resolution).

In conclusion, our data did not support the protective role of HLA-B27 on SARS-CoV-2 infection, including occurrence, ability to assert immune control of novel coronavirus and COVID-19 severity. More studies are needed to understand the role of HLA-B27 and variants of SARS-CoV-2.